CN118135802B - Bridge road management and control system and method based on deep learning network - Google Patents
Bridge road management and control system and method based on deep learning network Download PDFInfo
- Publication number
- CN118135802B CN118135802B CN202410572048.3A CN202410572048A CN118135802B CN 118135802 B CN118135802 B CN 118135802B CN 202410572048 A CN202410572048 A CN 202410572048A CN 118135802 B CN118135802 B CN 118135802B
- Authority
- CN
- China
- Prior art keywords
- road condition
- value
- lane
- road
- data
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- 238000000034 method Methods 0.000 title claims abstract description 44
- 238000013135 deep learning Methods 0.000 title claims abstract description 34
- 230000002159 abnormal effect Effects 0.000 claims abstract description 47
- 238000011217 control strategy Methods 0.000 claims abstract description 30
- 238000013209 evaluation strategy Methods 0.000 claims abstract description 19
- 230000008859 change Effects 0.000 claims description 145
- 230000006870 function Effects 0.000 claims description 34
- 238000012549 training Methods 0.000 claims description 24
- 230000009467 reduction Effects 0.000 claims description 18
- 230000015654 memory Effects 0.000 claims description 15
- 230000008569 process Effects 0.000 claims description 15
- 238000012216 screening Methods 0.000 claims description 12
- 230000006399 behavior Effects 0.000 claims description 9
- 238000012795 verification Methods 0.000 claims description 7
- 238000012360 testing method Methods 0.000 claims description 6
- 238000012545 processing Methods 0.000 claims description 5
- 238000012854 evaluation process Methods 0.000 claims description 4
- 238000011478 gradient descent method Methods 0.000 claims description 4
- 230000007787 long-term memory Effects 0.000 claims description 4
- 238000010606 normalization Methods 0.000 claims description 4
- 230000008901 benefit Effects 0.000 abstract description 5
- 230000009286 beneficial effect Effects 0.000 abstract description 2
- 238000007726 management method Methods 0.000 description 74
- 238000005457 optimization Methods 0.000 description 7
- 230000000694 effects Effects 0.000 description 5
- 238000005516 engineering process Methods 0.000 description 5
- 206010039203 Road traffic accident Diseases 0.000 description 4
- 238000002790 cross-validation Methods 0.000 description 4
- 238000012544 monitoring process Methods 0.000 description 4
- 230000003068 static effect Effects 0.000 description 4
- 230000009471 action Effects 0.000 description 3
- 230000000737 periodic effect Effects 0.000 description 3
- 238000013459 approach Methods 0.000 description 2
- 230000006403 short-term memory Effects 0.000 description 2
- 238000004458 analytical method Methods 0.000 description 1
- 238000007405 data analysis Methods 0.000 description 1
- 238000013499 data model Methods 0.000 description 1
- 230000007423 decrease Effects 0.000 description 1
- 230000003247 decreasing effect Effects 0.000 description 1
- 238000013136 deep learning model Methods 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 208000037265 diseases, disorders, signs and symptoms Diseases 0.000 description 1
- 238000011156 evaluation Methods 0.000 description 1
- 238000009472 formulation Methods 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 230000003993 interaction Effects 0.000 description 1
- 230000007774 longterm Effects 0.000 description 1
- 238000012423 maintenance Methods 0.000 description 1
- 239000000203 mixture Substances 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000003062 neural network model Methods 0.000 description 1
- 230000004044 response Effects 0.000 description 1
- 238000007619 statistical method Methods 0.000 description 1
- 238000012546 transfer Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G08—SIGNALLING
- G08G—TRAFFIC CONTROL SYSTEMS
- G08G1/00—Traffic control systems for road vehicles
- G08G1/01—Detecting movement of traffic to be counted or controlled
- G08G1/0104—Measuring and analyzing of parameters relative to traffic conditions
- G08G1/0125—Traffic data processing
- G08G1/0129—Traffic data processing for creating historical data or processing based on historical data
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/044—Recurrent networks, e.g. Hopfield networks
- G06N3/0442—Recurrent networks, e.g. Hopfield networks characterised by memory or gating, e.g. long short-term memory [LSTM] or gated recurrent units [GRU]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
- G06N3/084—Backpropagation, e.g. using gradient descent
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
- G06N3/092—Reinforcement learning
-
- G—PHYSICS
- G08—SIGNALLING
- G08B—SIGNALLING OR CALLING SYSTEMS; ORDER TELEGRAPHS; ALARM SYSTEMS
- G08B31/00—Predictive alarm systems characterised by extrapolation or other computation using updated historic data
-
- G—PHYSICS
- G08—SIGNALLING
- G08G—TRAFFIC CONTROL SYSTEMS
- G08G1/00—Traffic control systems for road vehicles
- G08G1/01—Detecting movement of traffic to be counted or controlled
- G08G1/0104—Measuring and analyzing of parameters relative to traffic conditions
- G08G1/0125—Traffic data processing
- G08G1/0133—Traffic data processing for classifying traffic situation
-
- G—PHYSICS
- G08—SIGNALLING
- G08G—TRAFFIC CONTROL SYSTEMS
- G08G1/00—Traffic control systems for road vehicles
- G08G1/01—Detecting movement of traffic to be counted or controlled
- G08G1/0104—Measuring and analyzing of parameters relative to traffic conditions
- G08G1/0137—Measuring and analyzing of parameters relative to traffic conditions for specific applications
- G08G1/0145—Measuring and analyzing of parameters relative to traffic conditions for specific applications for active traffic flow control
-
- G—PHYSICS
- G08—SIGNALLING
- G08G—TRAFFIC CONTROL SYSTEMS
- G08G1/00—Traffic control systems for road vehicles
- G08G1/01—Detecting movement of traffic to be counted or controlled
- G08G1/052—Detecting movement of traffic to be counted or controlled with provision for determining speed or overspeed
-
- G—PHYSICS
- G08—SIGNALLING
- G08G—TRAFFIC CONTROL SYSTEMS
- G08G1/00—Traffic control systems for road vehicles
- G08G1/065—Traffic control systems for road vehicles by counting the vehicles in a section of the road or in a parking area, i.e. comparing incoming count with outgoing count
-
- G—PHYSICS
- G08—SIGNALLING
- G08G—TRAFFIC CONTROL SYSTEMS
- G08G1/00—Traffic control systems for road vehicles
- G08G1/09—Arrangements for giving variable traffic instructions
- G08G1/0962—Arrangements for giving variable traffic instructions having an indicator mounted inside the vehicle, e.g. giving voice messages
- G08G1/0968—Systems involving transmission of navigation instructions to the vehicle
- G08G1/096805—Systems involving transmission of navigation instructions to the vehicle where the transmitted instructions are used to compute a route
- G08G1/096827—Systems involving transmission of navigation instructions to the vehicle where the transmitted instructions are used to compute a route where the route is computed onboard
-
- G—PHYSICS
- G08—SIGNALLING
- G08G—TRAFFIC CONTROL SYSTEMS
- G08G1/00—Traffic control systems for road vehicles
- G08G1/09—Arrangements for giving variable traffic instructions
- G08G1/0962—Arrangements for giving variable traffic instructions having an indicator mounted inside the vehicle, e.g. giving voice messages
- G08G1/0968—Systems involving transmission of navigation instructions to the vehicle
- G08G1/096833—Systems involving transmission of navigation instructions to the vehicle where different aspects are considered when computing the route
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02T—CLIMATE CHANGE MITIGATION TECHNOLOGIES RELATED TO TRANSPORTATION
- Y02T10/00—Road transport of goods or passengers
- Y02T10/10—Internal combustion engine [ICE] based vehicles
- Y02T10/40—Engine management systems
Landscapes
- Physics & Mathematics (AREA)
- Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Mathematical Physics (AREA)
- Computing Systems (AREA)
- Biophysics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- General Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Biomedical Technology (AREA)
- General Engineering & Computer Science (AREA)
- Artificial Intelligence (AREA)
- Computational Linguistics (AREA)
- Software Systems (AREA)
- Health & Medical Sciences (AREA)
- Analytical Chemistry (AREA)
- Chemical & Material Sciences (AREA)
- Radar, Positioning & Navigation (AREA)
- Remote Sensing (AREA)
- Emergency Management (AREA)
- Business, Economics & Management (AREA)
- Traffic Control Systems (AREA)
Abstract
The invention discloses a bridge road management and control system and method based on a deep learning network, comprising a data acquisition unit: the method comprises the steps that real-time road condition data acquisition is conducted on a target bridge road section in a unit management and control period; road condition prediction unit: setting the road condition of the target bridge road section into a plurality of abnormal road condition categories by using a road evaluation strategy, and iteratively predicting the road condition by using a deep learning network to the real-time road condition data of the target bridge road section; and the judging and executing unit: and implementing an emergency management and control strategy for all bridge sections meeting the abnormal road condition category. Compared with the prior art, the method has the advantages that the real-time acquisition is carried out on the vehicle and lane data on the bridge road, the deep learning network is used for iterating the real-time road condition data, the early warning deployment is carried out on the abnormal condition of the bridge road in advance by utilizing the predicted road condition, the real-time performance and the accuracy of the bridge road management and control are improved, and the method has the advantages and the beneficial effects of improving the prediction and the management and control capability of the bridge road traffic.
Description
Technical Field
The invention relates to the technical field of management and control, in particular to a bridge road management and control system and method based on a deep learning network.
Background
In the field of bridge road management and control, the prior art mainly comprises traditional traffic signal control, road condition monitoring systems, intelligent traffic management systems and the like. The intelligent traffic prediction and optimization algorithm is adopted in the technologies, and the problems of traffic jam and the like are prejudged in advance through data analysis and model prediction, so that the traffic efficiency and the road safety are improved. However, in bridge roads with longer length, the optimization of the traditional traffic management system on the bridge traffic capacity is often based on static traffic flow models and rules, the static traffic flow models and rules often cannot comprehensively consider traffic flow density change factors, so that the optimization effect is limited, comprehensive consideration on real-time traffic flow data and bridge structure states is lacking, the special conditions of the bridge roads are not fully considered, and the traffic capacity of the bridge roads cannot be fully predicted and optimized, so that road conditions are difficult to predict and road traffic is dredged in time when the number of vehicles is large.
Disclosure of Invention
The invention provides a bridge road management and control system and method based on a deep learning network, which solve the problem of poor road management and control capability caused by insufficient dynamic prediction of road vehicles by the existing bridge road management technology.
The invention is realized by the following technical scheme:
a bridge road management and control system based on a deep learning network, the system comprising:
A data acquisition unit: setting unit control periods for different bridge sections of a target bridge road, and collecting real-time road condition data of the target bridge section in the unit control periods, wherein the real-time road condition data comprise speed data of each vehicle and lane data of each lane, and recording the real-time road condition data based on ordinal numbers of the unit control periods;
Road condition prediction unit: setting a road evaluation strategy based on the real-time road condition data, setting the road condition of the target bridge road section into a plurality of abnormal road condition categories by using the road evaluation strategy, iterating the real-time road condition data of the target bridge road section by using a deep learning network to evaluate the road condition change tendency of the current bridge road section, and setting an iterated output value as predicted road condition data;
And the judging and executing unit: and setting emergency management and control strategies for each abnormal road condition category, screening all predicted road condition data conforming to the abnormal road condition category by using a road evaluation strategy, setting the predicted road condition data as marking data, screening out bridge road sections containing at least one marking data, and implementing all emergency management and control strategies conforming to the marking data aiming at all the screened bridge road sections.
In bridge roads with longer length, the optimization of the traditional traffic management system on the bridge traffic capacity is often based on a static traffic flow model and a rule, the static traffic flow model and the rule often cannot comprehensively consider traffic flow density change factors, so that the optimization effect is limited, the comprehensive consideration on real-time traffic flow data and bridge structure states is lacking, the dynamic optimization on the bridge traffic capacity cannot be realized, road conditions are difficult to predict when the number of vehicles is large, road traffic is dredged in time, and bridge bearing maintenance is also not facilitated. Based on the method, the invention provides a bridge road management and control system and method based on a deep learning network, which solve the problem that the dynamic optimization traffic capacity of the road is insufficient in the existing bridge road management technology.
Further, the process of setting the road assessment policy based on the real-time road condition data includes: the vehicle speed data comprises a deceleration frequency value, and the lane data comprises a first lane change value and a second lane change value; the deceleration frequency value represents the number of times of deceleration behaviors of each vehicle, and the first lane change value and the second lane change value are the number of times that the vehicle on each lane leaves the lane and enters the lane from the side at each time; recording the time point of each deceleration frequency value, the time point of the lane change completion of the first lane change value and the second lane change value on a target bridge section in a unit management and control period; and setting the abnormal road condition category based on the deceleration frequency value, the time point of the first lane change value and the second lane change value and the difference value.
Further, setting a deceleration critical value for the deceleration frequency value, and setting a lane change critical value for the first lane change value and the second lane change value; the abnormal road condition category includes:
first anomaly category: the speed reduction frequency value of the lane existing in the target bridge section is higher than a speed reduction critical value, and the first lane changing value and the second lane changing value of all lanes are not higher than the lane changing critical value;
Second anomaly category: the speed reduction frequency value of the lane existing in the target bridge section is higher than the speed reduction critical value, and the first lane changing value and the second lane changing value of the lane exist and are higher than the lane changing critical value;
Third anomaly category: the speed reduction frequency values of all lanes in the target bridge section are lower than a speed reduction critical value, and the first lane change value and the second lane change value of the lanes are higher than the lane change critical value;
Fourth anomaly category: setting a duty ratio threshold value for the duty ratio number of the deceleration frequency value and the first lane change value in the same lane, wherein the deceleration frequency value of the lane in the target bridge road section is higher than the deceleration critical value, the first lane change value and the second lane change value of the lane are higher than the lane change critical value, and the duty ratio number of the deceleration frequency value and the first lane change value on one lane reaches the duty ratio threshold value.
Further, the collecting form of the real-time road condition data includes: respectively recording vehicle speed data and lane data in the form of a continuous graph and a logic waveform graph by using a plane rectangular coordinate system, wherein the abscissa of the vehicle speed data and the lane data is set as the time length of a unit management and control period; the ordinate of the vehicle speed data represents the magnitude of the vehicle speed value; the ordinate of the lane data includes a first lane change value and a second lane change value.
Further, each unit management and control period is divided into three subcycles of a prediction period, a buffer period and an implementation period in sequence, real-time road condition data acquisition is carried out in the prediction period, and an emergency management and control strategy is started in the implementation period; and recording the predicted road condition data output in the predicted period and the actual road condition data in the implementation period as historical reference data, and updating the performance of the deep learning network and adjusting the time length distribution of the predicted period and the implementation period in the unit management period.
A bridge road management and control method based on a deep learning network comprises the following steps:
step S1: setting unit control periods for road control of the target bridge road for different bridge road sections respectively, and collecting real-time road condition data of the target bridge road in the unit control periods, wherein the real-time road condition data comprise speed data of each vehicle and lane data of each lane, and recording the real-time road condition data based on ordinal numbers of the unit control periods;
step S2: setting a road evaluation strategy based on real-time road condition data, setting reference road condition data for iterative reference, and classifying the real-time road condition data by using the road evaluation strategy to set the road condition of a target bridge road section into a plurality of abnormal road condition categories requiring emergency management and control processing according to the characteristics of management and control vehicles and management and control lanes;
Step S3: iterating real-time road condition data of a target bridge road section based on reference road condition data by using a deep learning network, predicting the running trend of all vehicles and the states of all lanes, evaluating the road condition change trend of the current bridge road section, and setting an iterated output value as predicted road condition data representing a predicted output result;
step S4: and setting emergency management and control strategies for each abnormal road condition category, screening all predicted road condition data conforming to the abnormal road condition category by using a road evaluation strategy, setting the predicted road condition data as marking data, screening out bridge road sections containing at least one marking data, and implementing all emergency management and control strategies conforming to the marking data aiming at all the screened bridge road sections.
Further, the road condition change tendency evaluation process is set as follows: iterating real-time road condition data of a target vehicle and a current driving lane by using a long-period and short-period memory network, wherein the contents comprise:
Constructing a long-period memory network, constructing a training set and a testing set, and arranging the deceleration frequency value of the target vehicle and the adjacent vehicles in front, the first lane change value and the second lane change value of the current lane of the target vehicle according to a time sequence and using the first lane change value and the second lane change value as input values of the long-period memory network; training the long-period memory network, minimizing the difference between the iteration result and the actual road condition data, outputting the iteration value as predicted road condition data after the iteration times are used up, and evaluating the prediction performance of the long-period memory network by using a cross verification method after each iteration is completed.
Further, the random gradient descent method is used for training the long-term and short-term memory network, and the method comprises the following steps:
Step A1: presetting a gradient initial value, iteration times and reference road condition data of gradient descent, and setting a deceleration frequency value of a target vehicle and a front adjacent vehicle, a first lane change value and a second lane change value of a current lane of the target vehicle as model input parameters;
Step A2: setting a mean square error loss function for the deceleration frequency value, setting a cross entropy loss function for the first lane change value and the second lane change value, and calculating a total loss function of gradient descent by weighting and summing the mean square error loss function and the cross entropy loss function;
step A3: calculating a gradient value of the total loss function according to the reference road condition data, iteratively outputting real-time road condition data of the target vehicle according to the gradient value, if the loss function is reduced, retaining the real-time road condition data output by the current iteration, otherwise, continuing to use the output value of the previous model until the iteration times are reached, and setting the output value reaching the iteration times as predicted road condition data.
Further, the time sequence sorting process includes: and adding time stamps into all deceleration frequency values of all vehicles in the real-time road condition data, all first lane changing values and second lane changing values of the current lane of the target vehicle for constructing a space-time characteristic sequence aiming at the target bridge road section in a unit management and control period, wherein each characteristic vector in the space-time characteristic sequence is formed by splicing at least one item of real-time road condition data and corresponding time stamps thereof in a data normalization mode.
Compared with the prior art, the method and the device have the advantages that the real-time acquisition is carried out on the vehicle and lane data on the bridge road, the deep learning network is used for iterating the real-time road condition data, the predicted road conditions of a plurality of abnormal road condition categories are generated, the predicted road conditions are utilized to early warn and deploy the abnormal conditions of the bridge road in advance, the real-time performance and the accuracy of bridge road management and control are improved, and the method and the device have the advantages and the beneficial effects of improving the prediction and management and control capability of the bridge road traffic.
Drawings
The accompanying drawings, which are included to provide a further understanding of embodiments of the application and are incorporated in and constitute a part of this specification, illustrate embodiments of the application and together with the description serve to explain the principles of the application. In the drawings:
FIG.1 is a schematic diagram of the structure of the present invention;
fig. 2 is a flow chart of the present invention.
Detailed Description
For the purpose of making apparent the objects, technical solutions and advantages of the present invention, the present invention will be further described in detail with reference to the following examples and the accompanying drawings, wherein the exemplary embodiments of the present invention and the descriptions thereof are for illustrating the present invention only and are not to be construed as limiting the present invention.
As shown in fig. 1, the bridge road management and control system based on the deep learning network of the present invention includes:
A data acquisition unit: setting unit control periods for different bridge sections of a target bridge road, and collecting real-time road condition data of the target bridge section in the unit control periods, wherein the real-time road condition data comprise speed data of each vehicle and lane data of each lane, and recording the real-time road condition data based on ordinal numbers of the unit control periods;
Road condition prediction unit: setting a road evaluation strategy based on the real-time road condition data, setting the road condition of the target bridge road section into a plurality of abnormal road condition categories by using the road evaluation strategy, iterating the real-time road condition data of the target bridge road section by using a deep learning network to evaluate the road condition change tendency of the current bridge road section, and setting an iterated output value as predicted road condition data;
And the judging and executing unit: and setting emergency management and control strategies for each abnormal road condition category, screening all predicted road condition data conforming to the abnormal road condition category by using a road evaluation strategy, setting the predicted road condition data as marking data, screening out bridge road sections containing at least one marking data, and implementing all emergency management and control strategies conforming to the marking data aiming at all the screened bridge road sections.
The bridge road section is a component section part of the target bridge road, namely, the purpose is to conduct road management and control on the target bridge road in a sectional mode, so that the management and control efficiency and the resource utilization rate are improved, and the system load is reduced. The unit management and control period is used for limiting the starting time of the bridge road management and control system, so that the resource consumption and the periodic management of the system are controlled conveniently. In the data acquisition unit, the length of the unit management cycle should be short enough to capture real-time changes in traffic flow, especially in the case of rapid changes in traffic flow during peak hours. Thus, the period length may vary depending on traffic flow conditions for different periods. In a specific implementation, the real-time road condition data may use sensor technology and video surveillance vision technology to obtain status data of each vehicle and each lane. Recording the ordinal number based on the unit management and control period for accumulating historical data for a system, and analyzing long-term trend, periodic change and occurrence of abnormal events for a target bridge section; by analyzing the historical data, the traffic flow characteristics under different time periods, different seasons or different weather conditions can be found, so that the management and control strategy is adjusted, the management and control strategy is more suitable for actual conditions, and the management and control effect is improved. The real-time road condition data is analyzed in an iterative mode on the target bridge road section through the deep learning network of the road condition prediction unit, the system can predict the road condition change trend of the current bridge road section, and the emergency management and control strategy can be correspondingly implemented by the judging and executing unit. In the same bridge road section in the unit management and control period, a plurality of predicted road condition data conforming to the abnormal road condition category label may exist, and the abnormal road condition categories conforming to the plurality of predicted road condition data may be the same or different. In order to ensure that the potential abnormal road conditions can be perfectly processed, even if the types of the abnormal road conditions which are met by the plurality of predicted road condition data are the same, all emergency management and control strategies which are met by the marking data are required to be implemented in specific implementation, namely, the potential abnormal road conditions can be provided with sufficient emergency processing resources, and the complete management and control of the bridge road, the target bridge road section of the bridge road is ensured. In the specific implementation, the real-time road condition data, the predicted road condition data and the execution condition of the emergency management and control strategy can be visually displayed, wherein the real-time traffic flow is displayed by using lines with different thickness, and the real-time vehicle speed distribution is displayed by using a thermal histogram; the curve trend graph is used for displaying the predicted road condition change and highlighting the possible traffic problem areas.
Further, as a possible implementation manner, the process of setting the road evaluation policy based on the real-time road condition data includes: the vehicle speed data comprises a deceleration frequency value, and the lane data comprises a first lane change value and a second lane change value; the deceleration frequency value represents the number of times of deceleration behaviors of each vehicle, and the first lane change value and the second lane change value are the number of times that the vehicle on each lane leaves the lane and enters the lane from the side at each time; recording the time point of each deceleration frequency value, the time point of the lane change completion of the first lane change value and the second lane change value on a target bridge section in a unit management and control period; and setting the abnormal road condition category based on the deceleration frequency value, the time point of the first lane change value and the second lane change value and the difference value. The deceleration frequency value is the number of times that each vehicle has deceleration behavior after entering a monitored and collected target bridge road section; the first lane change value and the second lane change value represent the number of times the vehicle laterally leaves the lane and laterally enters the lane on each lane, respectively. The time point of each deceleration frequency value and the time point of the lane change completion of the first lane change value and the second lane change value are recorded, so that the deceleration behavior and the lane change condition of each vehicle can be monitored, the specific time point is recorded, the method can be used for analyzing and evaluating the driving behavior, and a basis is provided for the formulation and adjustment of the management and control strategy. In a specific implementation, an action duration range can be set along two sides of a time axis according to time points of a deceleration frequency value, a first lane change value and a second lane change value, when the deceleration frequency value of a first vehicle coincides with the first lane change value of a second vehicle or the action duration range of the deceleration frequency value of the first vehicle coincides with the action duration range of the second lane change value of the second vehicle, namely, it is determined that motion interference exists between the first vehicle and the second vehicle, when the number of the motion interference of the same bridge section in the same unit management and control period reaches a preset threshold, the prediction implementation frequency of abnormal road condition categories of the bridge section is increased, and the time period of the unit management and control period and the monitored real-time road condition data record are used as historical data for periodic analysis.
Further, as a possible implementation manner, a deceleration critical value is set for the deceleration frequency value, and a lane change critical value is set for the first lane change value and the second lane change value; the abnormal road condition category includes:
first anomaly category: the speed reduction frequency value of the lane existing in the target bridge section is higher than a speed reduction critical value, and the first lane changing value and the second lane changing value of all lanes are not higher than the lane changing critical value;
Second anomaly category: the speed reduction frequency value of the lane existing in the target bridge section is higher than the speed reduction critical value, and the first lane changing value and the second lane changing value of the lane exist and are higher than the lane changing critical value;
Third anomaly category: the speed reduction frequency values of all lanes in the target bridge section are lower than a speed reduction critical value, and the first lane change value and the second lane change value of the lanes are higher than the lane change critical value;
Fourth anomaly category: setting a duty ratio threshold value for the duty ratio number of the deceleration frequency value and the first lane change value in the same lane, wherein the deceleration frequency value of the lane in the target bridge road section is higher than the deceleration critical value, the first lane change value and the second lane change value of the lane are higher than the lane change critical value, and the duty ratio number of the deceleration frequency value and the first lane change value on one lane reaches the duty ratio threshold value.
Since all lanes are monitored simultaneously in the target bridge section, the total value of the first lane change values and the total value of the second lane change values are approximately equal in the target bridge section. The first abnormal category represents traffic congestion on the bridge, and vehicles need to be braked frequently but have variable track behaviors, which may be caused by congestion due to large traffic flow. The emergency management and control strategy aiming at the first abnormal category can be to optimize or dispatch traffic signals to guide traffic of traffic police, guide vehicles to split or change driving routes so as to relieve congestion. The second abnormal category indicates that the traffic condition on the bridge is complex, the vehicle brakes frequently and changes lanes frequently, and traffic jam or traffic disorder caused by large traffic flow can exist. The emergency management and control strategy for the second abnormal category may be to add temporary traffic restrictions for the target bridge segment: temporary traffic restriction measures are implemented on specific road segments or areas to reduce the occurrence of traffic upsets. The third abnormal category indicates that vehicles on the bridge road section frequently perform lane changing behavior, but frequent braking is not needed, and the traffic of each lane is not blocked and still keeps certain running smoothness, which is probably caused by frequent lane changing conditions caused by road structures such as intersections, ramps and the like. The emergency management and control strategy aiming at the third abnormal category can strengthen the supervision and management of intersections and ramps and reduce the frequent lane changing conditions caused by the structures of the intersections and the ramps. The fourth abnormal category indicates abnormal traffic conditions on the bridge, vehicles are braked frequently and changed frequently, and serious conditions such as traffic accidents, vehicle detention due to reasons and the like possibly exist, so that the running of a single lane is stopped, and management and control measures are needed to be adopted in time. The emergency control strategy aiming at the fourth abnormal category can immediately start a traffic accident emergency plan, send a rescue vehicle to the accident site, and rapidly treat and rescue the traffic accident emergency plan so as to reduce the influence of the accident on the traffic; meanwhile, traffic restriction is implemented according to traffic conditions, traffic guiding measures are adopted to guide vehicles to detour or transfer traffic so as to reduce congestion and accidents. The setting modes of the deceleration critical value, the lane change critical value and the duty ratio threshold value can be set as follows: determining a reasonable range of the duty ratio threshold value of the deceleration frequency value and the first lane change value in the same lane according to the past traffic data and experience; and calculating the frequency distribution condition of the simultaneous occurrence of the deceleration frequency value and the first lane change value on the same lane by using a statistical analysis method, and selecting a proper percentile or distribution interval as a duty ratio threshold according to a statistical result so as to ensure that the threshold can cover most abnormal conditions.
Further, as a possible implementation manner, the collecting form of the real-time road condition data includes: respectively recording vehicle speed data and lane data in the form of a continuous graph and a logic waveform graph by using a plane rectangular coordinate system, wherein the abscissa of the vehicle speed data and the lane data is set as the time length of a unit management and control period; the ordinate of the vehicle speed data represents the magnitude of the vehicle speed value; the ordinate of the lane data includes a first lane change value and a second lane change value. The continuous graph is used for recording the vehicle speed data, the abscissa is used for controlling the time length of the period, and the ordinate represents the magnitude of the vehicle speed value. The continuous graph can clearly show the change trend of the vehicle speed along with time, including the vehicle speed fluctuation at peak time, the congestion condition, the traffic flow concentration and the like. The logic relationship of lane change conditions, such as the change trend of the first lane change value and the second lane change value and the comparison thereof, can be intuitively displayed by using the logic waveform chart to record the lane data, and is helpful for identifying the frequent lane change or the condition of uneven traffic flow distribution. Through the acquisition form, a user can intuitively know the conditions of vehicle speed and lane change so as to take corresponding traffic control measures in time.
Further, as a feasible implementation mode, each unit management and control period is divided into a prediction period, a buffer period and an implementation period in sequence, real-time road condition data acquisition is carried out in the prediction period, and an emergency management and control strategy is started in the implementation period; and recording the predicted road condition data output in the predicted period and the actual road condition data in the implementation period as historical reference data, and updating the performance of the deep learning network and adjusting the time length distribution of the predicted period and the implementation period in the unit management period. The management and control period is divided into different sub-periods, so that the system can adjust the execution time of the data acquisition, prediction and management and control strategy more flexibly according to the actual situation. The process of data monitoring and acquisition is placed in a prediction period, and the training and prediction of the deep learning network can be performed by utilizing historical data and real-time data. The implementation period is divided into a buffer period and an implementation period, preparation can be carried out in the prediction period according to the prediction result, and the buffer period provides response time, so that the process of data monitoring and acquisition is indirectly advanced, emergency management and control strategies can be started more quickly in the implementation period, road condition changes can be dealt with in time, and the occurrence probability of traffic accidents and congestion is reduced. By recording the predicted road condition data output in the predicted period and the actual road condition data in the implementation period as historical reference data, the effect of the management and control strategy and the change trend of the road condition can be monitored in real time, and important basis is provided for the performance update of the deep learning network and the time length distribution of the predicted period and the implementation period in the unit management and control period. This helps optimize the management and control strategy, improves traffic management's efficiency and precision.
Further, as another embodiment, as shown in fig. 2, a bridge road management method based on a deep learning network, the method includes:
step S1: setting unit control periods for road control of the target bridge road for different bridge road sections respectively, and collecting real-time road condition data of the target bridge road in the unit control periods, wherein the real-time road condition data comprise speed data of each vehicle and lane data of each lane, and recording the real-time road condition data based on ordinal numbers of the unit control periods;
step S2: setting a road evaluation strategy based on real-time road condition data, setting reference road condition data for iterative reference, and classifying the real-time road condition data by using the road evaluation strategy to set the road condition of a target bridge road section into a plurality of abnormal road condition categories requiring emergency management and control processing according to the characteristics of management and control vehicles and management and control lanes;
Step S3: iterating real-time road condition data of a target bridge road section based on reference road condition data by using a deep learning network, predicting the running trend of all vehicles and the states of all lanes, evaluating the road condition change trend of the current bridge road section, and setting an iterated output value as predicted road condition data representing a predicted output result;
step S4: and setting emergency management and control strategies for each abnormal road condition category, screening all predicted road condition data conforming to the abnormal road condition category by using a road evaluation strategy, setting the predicted road condition data as marking data, screening out bridge road sections containing at least one marking data, and implementing all emergency management and control strategies conforming to the marking data aiming at all the screened bridge road sections.
Further, as a possible embodiment, the evaluation process of the road condition change tendency is set as: iterating real-time road condition data of a target vehicle and a current driving lane by using a long-period and short-period memory network, wherein the contents comprise:
Constructing a long-period memory network, constructing a training set and a testing set, and arranging the deceleration frequency value of the target vehicle and the adjacent vehicles in front, the first lane change value and the second lane change value of the current lane of the target vehicle according to a time sequence and using the first lane change value and the second lane change value as input values of the long-period memory network; training the long-period memory network, minimizing the difference between the iteration result and the actual road condition data, outputting the iteration value as predicted road condition data after the iteration times are used up, and evaluating the prediction performance of the long-period memory network by using a cross verification method after each iteration is completed.
In a long and short term memory network, the training set contains data samples for training a model. These samples are typically pre-processed and feature engineered to accommodate the input format and requirements of the deep learning model. The data samples of the training set cover various road conditions, such as different vehicle speeds, different lane changes, corresponding time point data, and the like. The test set also contains real-time road condition data samples, but unlike the training set, the data samples of the test set are setup data that the model has not seen during the training process. The data samples are derived from practically reachable but not yet occurring data estimated by the system according to the existing real-time road condition data, and are used for evaluating the performance of the model on unknown data so as to verify the generalization capability and performance of the model. The deceleration frequency value in the time series may be the number of decelerations of the target vehicle at a certain time point, or may be the number of common decelerations of the target vehicle and the adjacent vehicle in a period of time. Collecting the deceleration frequency values of the target vehicle and the front adjacent vehicle can provide more comprehensive road condition information. The deceleration behavior of adjacent vehicles may affect the travel of the target vehicle, and interactions between vehicles may be more pronounced, especially in high density traffic situations. Therefore, the actual traffic condition of the current road can be reflected more accurately by collecting the deceleration frequency values of the adjacent vehicles. And collecting a first lane change value and a second lane change value of the current lane of the target vehicle, wherein the first lane change value and the second lane change value reflect the frequency of the target vehicle for changing lanes on the current lane.
More, the minimized difference between the iteration result and the actual road condition data means that the model gradually adjusts the parameters thereof in the training process, so that the prediction capability of the model on the actual road condition gradually approaches to the actual condition. By continuously reducing the difference between the prediction result and the actual data, the accuracy and the prediction precision of the model can be improved, so that the model can better capture the change trend and the characteristics of road conditions. By minimizing the difference in the training process, the deviation and error of the model in the prediction process can be reduced, the model not only has good performance on training data, but also can be better adapted to unseen data, and has stronger generalization capability. The cross-validation method can evaluate the performance of the model on unseen data, rather than relying on a single training set and test set, and helps to fully understand the predictive effect of the model on new data.
In a specific implementation, as a possible implementation, preferably, the cross-validation method may use a K-fold cross-validation method: and designing and constructing an LSTM neural network model, wherein the model takes real-time road condition data and a time stamp as input to predict the road condition data at the future moment. In the process of cross verification, a training set containing real-time road condition data is divided into K mutually exclusive subsets (usually, K is 5 or 10), one subset is taken as a verification set at a time, and the remaining K-1 subsets are taken as training sets; model training is carried out by using K-1 subsets for each verification, then the remaining 1 subsets are used for verification, K times are repeated, each subset is ensured to serve as a verified set, different verified sets are used for each time, K models can be obtained, and performance indexes (such as accuracy, loss value, mean Square Error (MSE), mean Absolute Error (MAE) and the like) of the models are calculated and processed averagely; according to the cross-validation result, super parameters of the LSTM network, such as network structure, learning rate, iteration number and the like, can be optimized, so that the prediction performance of the model is further improved.
Further, as a possible implementation, training the long-term memory network by using a random gradient descent method includes:
Step A1: presetting a gradient initial value, iteration times and reference road condition data of gradient descent, and setting a deceleration frequency value of a target vehicle and a front adjacent vehicle, a first lane change value and a second lane change value of a current lane of the target vehicle as model input parameters;
Step A2: setting a mean square error loss function for the deceleration frequency value, setting a cross entropy loss function for the first lane change value and the second lane change value, and calculating a total loss function of gradient descent by weighting and summing the mean square error loss function and the cross entropy loss function;
step A3: calculating a gradient value of the total loss function according to the reference road condition data, iteratively outputting real-time road condition data of the target vehicle according to the gradient value, if the loss function is reduced, retaining the real-time road condition data output by the current iteration, otherwise, continuing to use the output value of the previous model until the iteration times are reached, and setting the output value reaching the iteration times as predicted road condition data.
The gradient initial value is an initial value set when a random gradient descent method is used for starting the gradient descent process. In each iteration, the gradient represents the rate of change of the loss function relative to the model parameters, guiding the update direction of the model parameters. The reference road condition data is used for calculating a total loss function, and the purpose of the reference road condition data is to measure the difference between the model predicted value and the actual observed value. In each iteration, the gradient value of the loss function is calculated according to the reference road condition data, which means that the change rate of the loss function with respect to the model parameters represents the direction in which the loss function increases fastest in the parameter space, so that the opposite direction of the gradient is the direction in which the loss function decreases, and by updating the parameters along the opposite direction of the gradient, the loss function can be gradually decreased, and the model gradually approaches to the optimal solution. And setting a cross entropy loss function for the first lane change value and the second lane change value, and measuring the difference between the actual lane change condition and the lane change condition predicted by the model. The mean square error loss function and the cross entropy loss function are respectively used, and are set according to the discrete and single logic statistical characteristics of the deceleration frequency value, the first lane change value and the second lane change value based on the characteristics of real-time road condition data. And weighted summation as a total loss function can make the model have good adaptability in different types of tasks. For example, in predicting the deceleration frequency value, the mean square error loss function can better reflect the distance between the predicted value and the actual value, and the cross entropy loss function can be used to deal with classification problems, such as prediction of lane change conditions. The setting mode of the reference road condition data can be used for taking actual road condition data collected in a past period of time as a reference, such as vehicle speed data, lane change data and the like, or real-time road condition data which is collected by real-time monitoring equipment and has no abnormal road condition category in normal running is used as the reference road condition data.
Further, as a possible implementation manner, the time sequence sorting process includes: and adding time stamps into all deceleration frequency values of all vehicles in the real-time road condition data, all first lane changing values and second lane changing values of the current lane of the target vehicle for constructing a space-time characteristic sequence aiming at the target bridge road section in a unit management and control period, wherein each characteristic vector in the space-time characteristic sequence is formed by splicing at least one item of real-time road condition data and corresponding time stamps thereof in a data normalization mode.
The time correlation of the real-time road condition data can be captured by adding the time stamp. Because traffic conditions may change over time, recording the acquisition time of each data point can help the model better understand and predict traffic conditions for different time periods. The addition of the time stamp causes the data to become time series data, i.e., a time series of time series features. This helps the model learn the timing relationships of the data, e.g., the model can identify at which time period traffic congestion begins to intensify, or how certain traffic events affect road conditions over time. The time-space characteristic sequence is a sequence data structure and is used for representing road condition data in a period of time. Each time point in the time-space characteristic sequence has a characteristic vector, and the characteristic vector comprises real-time road condition data and corresponding time stamps in a period of time. The data normalization process is to uniformly scale the data to a specific range, and eliminate dimension differences among different features, so that model training is more stable, convergence speed is increased, and meanwhile, the model training is prevented from being greatly influenced by certain features. When the feature vector is constructed, the real-time road condition data and the time stamp are spliced together in a certain mode to form a complete feature vector. The splicing mode can be that the data and the time stamp are connected together in sequence simply when the method is specifically applied, or that the splicing is performed after the data processing.
The foregoing description of the embodiments has been provided for the purpose of illustrating the general principles of the invention, and is not meant to limit the scope of the invention, but to limit the invention to the particular embodiments, and any modifications, equivalents, improvements, etc. that fall within the spirit and principles of the invention are intended to be included within the scope of the invention.
Claims (8)
1. Bridge road management and control system based on deep learning network, characterized by that, this system includes:
A data acquisition unit: setting unit control periods for different bridge sections of a target bridge road, and collecting real-time road condition data of the target bridge section in the unit control periods, wherein the real-time road condition data comprise speed data of each vehicle and lane data of each lane, and recording the real-time road condition data based on ordinal numbers of the unit control periods;
Road condition prediction unit: setting a road evaluation strategy based on the real-time road condition data, setting the road condition of the target bridge road section into a plurality of abnormal road condition categories by using the road evaluation strategy, iterating the real-time road condition data of the target bridge road section by using a deep learning network to evaluate the road condition change tendency of the current bridge road section, and setting an iterated output value as predicted road condition data;
And the judging and executing unit: setting an emergency management and control strategy corresponding to each abnormal road condition category, screening all predicted road condition data conforming to the abnormal road condition category by using a road evaluation strategy, setting the predicted road condition data as marking data, screening out bridge sections containing at least one marking data, and implementing all emergency management and control strategies conforming to the marking data aiming at all the screened bridge sections;
The evaluation process of the road condition change tendency is set as follows: arranging the deceleration frequency value of the target vehicle and the adjacent vehicles in front, the first lane changing value and the second lane changing value of the current lane of the target vehicle according to a time sequence and taking the time sequence as input values, and taking the iteration values as predicted road condition data to output;
The process of setting the road assessment strategy based on the real-time road condition data comprises the following steps: the vehicle speed data comprises a deceleration frequency value, and the lane data comprises a first lane change value and a second lane change value; the deceleration frequency value represents the number of times of deceleration behaviors of each vehicle, and the first lane change value and the second lane change value are the number of times that the vehicle on each lane leaves the lane and enters the lane from the side at each time; recording the time point of each deceleration frequency value, the time point of the lane change completion of the first lane change value and the second lane change value on a target bridge section in a unit management and control period; and setting the abnormal road condition category based on the deceleration frequency value, the time point of the first lane change value and the second lane change value and the difference value.
2. The bridge road control system based on the deep learning network according to claim 1, wherein a deceleration critical value is set for the deceleration frequency value, and a lane change critical value is set for the first lane change value and the second lane change value; the abnormal road condition category includes:
first anomaly category: the speed reduction frequency value of the lane existing in the target bridge section is higher than a speed reduction critical value, and the first lane changing value and the second lane changing value of all lanes are not higher than the lane changing critical value;
Second anomaly category: the speed reduction frequency value of the lane existing in the target bridge section is higher than the speed reduction critical value, and the first lane changing value and the second lane changing value of the lane exist and are higher than the lane changing critical value;
Third anomaly category: the speed reduction frequency values of all lanes in the target bridge section are lower than a speed reduction critical value, and the first lane change value and the second lane change value of the lanes are higher than the lane change critical value;
Fourth anomaly category: setting a duty ratio threshold value for the duty ratio number of the deceleration frequency value and the first lane change value in the same lane, wherein the deceleration frequency value of the lane in the target bridge road section is higher than the deceleration critical value, the first lane change value and the second lane change value of the lane are higher than the lane change critical value, and the duty ratio number of the deceleration frequency value and the first lane change value on one lane reaches the duty ratio threshold value.
3. The bridge road management and control system based on the deep learning network according to claim 1, wherein the acquisition form of the real-time road condition data comprises: respectively recording vehicle speed data and lane data in the form of a continuous graph and a logic waveform graph by using a plane rectangular coordinate system, wherein the abscissa of the vehicle speed data and the lane data is set as the time length of a unit management and control period; the ordinate of the vehicle speed data represents the magnitude of the vehicle speed value; the ordinate of the lane data includes a first lane change value and a second lane change value.
4. The bridge road management and control system based on the deep learning network according to claim 1, wherein each unit management and control period is divided into a prediction period, a buffer period and an implementation period in sequence, real-time road condition data acquisition is carried out in the prediction period, and an emergency management and control strategy is started in the implementation period; and recording the predicted road condition data output in the predicted period and the actual road condition data in the implementation period as historical reference data, and updating the performance of the deep learning network and adjusting the time length distribution of the predicted period and the implementation period in the unit management period.
5. A bridge road management and control method based on a deep learning network, based on the bridge road management and control system based on a deep learning network as set forth in any one of claims 1 to 4, characterized in that the method comprises:
step S1: setting unit control periods for road control of the target bridge road for different bridge road sections respectively, and collecting real-time road condition data of the target bridge road in the unit control periods, wherein the real-time road condition data comprise speed data of each vehicle and lane data of each lane, and recording the real-time road condition data based on ordinal numbers of the unit control periods;
step S2: setting a road evaluation strategy based on real-time road condition data, setting reference road condition data for iterative reference, and classifying the real-time road condition data by using the road evaluation strategy to set the road condition of a target bridge road section into a plurality of abnormal road condition categories requiring emergency management and control processing according to the characteristics of management and control vehicles and management and control lanes;
Step S3: iterating real-time road condition data of a target bridge road section based on reference road condition data by using a deep learning network, predicting the running trend of all vehicles and the states of all lanes, evaluating the road condition change trend of the current bridge road section, and setting an iterated output value as predicted road condition data representing a predicted output result;
step S4: and setting emergency management and control strategies for each abnormal road condition category, screening all predicted road condition data conforming to the abnormal road condition category by using a road evaluation strategy, setting the predicted road condition data as marking data, screening out bridge road sections containing at least one marking data, and implementing all emergency management and control strategies conforming to the marking data aiming at all the screened bridge road sections.
6. The method for managing and controlling bridge roads based on deep learning network according to claim 5, wherein the evaluation process of the road condition change tendency is set as follows: iterating real-time road condition data of a target vehicle and a current driving lane by using a long-period and short-period memory network, wherein the contents comprise:
Constructing a long-period memory network, constructing a training set and a testing set, and arranging the deceleration frequency value of the target vehicle and the adjacent vehicles in front, the first lane change value and the second lane change value of the current lane of the target vehicle according to a time sequence and using the first lane change value and the second lane change value as input values of the long-period memory network; training the long-period memory network, minimizing the difference between the iteration result and the actual road condition data, outputting the iteration value as predicted road condition data after the iteration times are used up, and evaluating the prediction performance of the long-period memory network by using a cross verification method after each iteration is completed.
7. The method for managing and controlling bridge roads based on deep learning network according to claim 6, wherein the training of the long-term memory network by using the random gradient descent method comprises the following steps:
Step A1: presetting a gradient initial value, iteration times and reference road condition data of gradient descent, and setting a deceleration frequency value of a target vehicle and a front adjacent vehicle, a first lane change value and a second lane change value of a current lane of the target vehicle as model input parameters;
Step A2: setting a mean square error loss function for the deceleration frequency value, setting a cross entropy loss function for the first lane change value and the second lane change value, and calculating a total loss function of gradient descent by weighting and summing the mean square error loss function and the cross entropy loss function;
step A3: calculating a gradient value of the total loss function according to the reference road condition data, iteratively outputting real-time road condition data of the target vehicle according to the gradient value, if the loss function is reduced, retaining the real-time road condition data output by the current iteration, otherwise, continuing to use the output value of the previous model until the iteration times are reached, and setting the output value reaching the iteration times as predicted road condition data.
8. The method for managing and controlling bridge roads based on deep learning network according to claim 6, wherein the time series arrangement process comprises: and adding time stamps into all deceleration frequency values of all vehicles in the real-time road condition data, all first lane changing values and second lane changing values of the current lane of the target vehicle for constructing a space-time characteristic sequence aiming at the target bridge road section in a unit management and control period, wherein each characteristic vector in the space-time characteristic sequence is formed by splicing at least one item of real-time road condition data and corresponding time stamps thereof in a data normalization mode.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202410572048.3A CN118135802B (en) | 2024-05-10 | 2024-05-10 | Bridge road management and control system and method based on deep learning network |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202410572048.3A CN118135802B (en) | 2024-05-10 | 2024-05-10 | Bridge road management and control system and method based on deep learning network |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN118135802A CN118135802A (en) | 2024-06-04 |
| CN118135802B true CN118135802B (en) | 2024-06-28 |
Family
ID=91240817
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202410572048.3A Active CN118135802B (en) | 2024-05-10 | 2024-05-10 | Bridge road management and control system and method based on deep learning network |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN118135802B (en) |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104933863A (en) * | 2015-06-02 | 2015-09-23 | 福建工程学院 | Method and system for recognizing abnormal segment of traffic road |
| CN109285349A (en) * | 2018-11-02 | 2019-01-29 | 中电海康集团有限公司 | Freeway traffic event detection method and early warning system under bus or train route cooperative surroundings |
Family Cites Families (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP4640121B2 (en) * | 2005-11-09 | 2011-03-02 | 住友電気工業株式会社 | Traffic information processing apparatus and traffic information processing method |
| JP4728315B2 (en) * | 2007-12-03 | 2011-07-20 | 住友電気工業株式会社 | Traffic volume calculation device, traffic volume calculation program, and traffic volume calculation method |
| JP6505349B2 (en) * | 2017-03-10 | 2019-04-24 | 三菱電機株式会社 | INFORMATION PROCESSING APPARATUS, INFORMATION PROCESSING METHOD, AND INFORMATION PROCESSING PROGRAM |
| JP2019168899A (en) * | 2018-03-23 | 2019-10-03 | パイオニア株式会社 | Congestion information determination device |
| GB2585243A (en) * | 2019-07-05 | 2021-01-06 | Valerann Ltd | Traffic event and road condition identification and classification |
| US11648938B2 (en) * | 2019-10-02 | 2023-05-16 | Toyota Motor North America, Inc. | Braking data mapping |
| CN114495486B (en) * | 2021-12-28 | 2023-06-09 | 江苏泰坦智慧科技有限公司 | Microcosmic traffic flow prediction system and microcosmic traffic flow prediction method based on hierarchical reinforcement learning |
| CN115050177B (en) * | 2022-05-11 | 2023-04-25 | 杭州像素元科技有限公司 | Expressway congestion prediction method based on space-time multi-head attention mechanism |
| CN115063054B (en) * | 2022-08-17 | 2022-11-29 | 深圳市索菱实业股份有限公司 | Internet of vehicles big data analysis method and server |
| CN116631186B (en) * | 2023-05-19 | 2024-05-28 | 东南大学 | Expressway traffic accident risk assessment method and system based on dangerous driving event data |
-
2024
- 2024-05-10 CN CN202410572048.3A patent/CN118135802B/en active Active
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104933863A (en) * | 2015-06-02 | 2015-09-23 | 福建工程学院 | Method and system for recognizing abnormal segment of traffic road |
| CN109285349A (en) * | 2018-11-02 | 2019-01-29 | 中电海康集团有限公司 | Freeway traffic event detection method and early warning system under bus or train route cooperative surroundings |
Also Published As
| Publication number | Publication date |
|---|---|
| CN118135802A (en) | 2024-06-04 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN112085947B (en) | Traffic jam prediction method based on deep learning and fuzzy clustering | |
| Kalair et al. | Anomaly detection and classification in traffic flow data from fluctuations in the flow–density relationship | |
| CN115148019A (en) | Early warning method and system based on holiday congestion prediction algorithm | |
| CN111179591B (en) | Road network traffic time sequence characteristic data quality diagnosis and restoration method | |
| CN115760484A (en) | Method, device and system for improving potential danger identification capability of power distribution station area and storage medium | |
| CN116597642A (en) | Traffic jam condition prediction method and system | |
| CN113159374A (en) | Data-driven urban traffic flow rate mode identification and real-time prediction early warning method | |
| CN117719385A (en) | Intelligent charging pile control system and method | |
| CN118041661A (en) | Abnormal network flow monitoring method, device and equipment based on deep learning and readable storage medium | |
| Siddique et al. | State-dependent self-adaptive sampling (SAS) method for vehicle trajectory data | |
| Elhenawy et al. | Automatic congestion identification with two-component mixture models | |
| Xu et al. | A deep learning approach for vehicle velocity prediction considering the influence factors of multiple lanes | |
| Han et al. | Transformer-based modeling of abnormal driving events for freeway crash risk evaluation | |
| CN118135802B (en) | Bridge road management and control system and method based on deep learning network | |
| CN117953693B (en) | Expressway traffic intelligent monitoring system based on video ai | |
| Kamal et al. | Early detection of a traffic flow breakdown in the freeway based on dynamical network markers | |
| Kim et al. | Reinforcement learning approach to develop variable speed limit strategy using vehicle data and simulations | |
| EP4060576A1 (en) | A method for identifying vehicle performance | |
| CN118172936B (en) | Intelligent abnormal traffic event detection method and system based on lane-level traffic parameters | |
| CN117633489B (en) | AI-based vehicle safety analysis method and system | |
| CN118711376B (en) | Traffic early warning method for intelligent traffic control system | |
| CN118314728A (en) | Traffic flow dynamic prediction and precision optimization method based on independent thermal coding | |
| CN113380032B (en) | Hierarchical clustering method-based highway congestion judgment method and device | |
| Mei et al. | Research on short-term urban traffic congestion based on fuzzy comprehensive evaluation and machine learning | |
| CN118865705A (en) | Intelligent traffic flow prediction method based on Bayesian algorithm |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |